[ATOM-vLLM] Add support for Qwen3.8

[kliuae]

Motivation

This PR adds support for Qwen3.8 in ATOM-vLLM mode.
Currently only tested with MXFP4 weights on single-node MI355 x8.

Technical Details

• Add and register text-only Qwen3_5
• Resolve ssm state dtype to dtype from model config unless otherwise set

Test Plan

lm_eval with gsm8k (single node MI355 x8, TP8)

Model: amd/Qwen3.8-2.4T-A95B-Quark-MXFP4

Server command

TP=8
export VLLM_ROCM_USE_AITER=1
export ATOM_ENABLE_QK_NORM_ROPE_CACHE_QUANT_FUSION=1
export ATOM_USE_CUSTOM_ALL_GATHER=0
export ATOM_FP8_BLOCKSCALE_WEIGHT_PRESHUFFLE=0
export GATED_DELTA_RULE_TRITON_AUTOTUNE=1

vllm serve amd/Qwen3.8-2.4T-A95B-Quark-MXFP4 \
    --host localhost \
    --port 8000 \
    --tensor-parallel-size "${TP}" \
    --kv-cache-dtype fp8 \
    --async-scheduling \
    --trust-remote-code \
    --max-num-batched-tokens 16384 \
    --max-model-len 16384 \
    --compilation-config '{"cudagraph_mode": "FULL_AND_PIECEWISE"}' \
    --gpu-memory-utilization 0.9 \
    --no-enable-prefix-caching

lm_eval

lm_eval --model local-completions \
        --model_args model=amd/Qwen3.8-2.4T-A95B-Quark-MXFP4,base_url=http://localhost:8000/v1/completions,num_concurrent=64,max_retries=3,tokenized_requests=False \
        --tasks gsm8k \
        --num_fewshot 5 \
        --gen_kwargs max_gen_toks=2048

Test Result

gsm8k

local-completions ({'model': 'amd/Qwen3.8-2.4T-A95B-Quark-MXFP4', 'base_url': 'http://localhost:8000/v1/completions', 'num_concurrent': 64, 'max_retries': 3, 'tokenized_requests': False}), gen_kwargs: ({'max_gen_toks': 2048}), limit: None, num_fewshot: 5, batch_size: 1

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
gsm8k	3	flexible-extract	5	exact_match	↑	0.9591	±	0.0055
		strict-match	5	exact_match	↑	0.9598	±	0.0054

Submission Checklist

• Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.

[github-actions]

🏷️ CI Guide

Runs automatically on every eligible PR before approval:

• ✅ Pre Checkin: Black, Ruff, catalog schema validation, non-GPU unit tests

Heavy model tests:

• ✅ Run after the PR is approved and Pre Checkin passes
• ✅ Run immediately when an approval review is submitted
• ✅ Can be requested before approval with labels

Label	Tests
ci:full	Run all heavy PR model tests: native ATOM, vLLM, and SGLang
ci:atom	Run native ATOM model accuracy tests
ci:vllm	Run ATOM vLLM OOT model accuracy tests
ci:sglang	Run ATOM SGLang model accuracy tests

Heavy jobs are skipped when the PR is not approved and no matching ci:* label is present.
Add labels via the sidebar or gh pr edit 1901 --add-label <label>